fix

Author: markvanderloo

pkg/DESCRIPTION

@@ -23,7 +23,7 @@ Description: Implements an approximate string matching version of R's native
  vectors representing generic sequences. This package is built for speed and
  runs in parallel by using 'openMP'. An API for C or C++ is exposed as well.
  Reference: MPJ van der Loo (2014) <doi:10.32614/RJ-2014-011>.
-Version: 0.9.12
+Version: 0.9.13
 Depends:
  R (>= 2.15.3)
 URL: https://github.com/markvanderloo/stringdist
@@ -32,4 +32,4 @@ Suggests:
  tinytest
 Imports: parallel
 Encoding: UTF-8
-RoxygenNote: 7.2.3
+RoxygenNote: 7.3.2

pkg/NEWS

@@ -1,3 +1,7 @@
+version 0.9.13
+- Fixed issue with zero-length strings in 'qgrams' (Thanks to Brian Ripley
+ for the notification and pointer to the origin of the problem)
+
 version 0.9.12
 - apparently R_xlen_t is long long int on CLANG/Windows and long int on gcc-13/debian
 

pkg/R/phonetic.R

@@ -17,7 +17,7 @@
 #' not be trusted. If non-ascii or non-printable ascii charcters are encountered, a warning 
 #' is emitted.
 #' 
-#' @seealso \code{\link{printable_ascii}}, \code{\link{stringdist-package}}
+#' @seealso \code{\link{printable_ascii}}
 #' 
 #' 
 #' @return

pkg/R/qgrams.R

@@ -23,12 +23,21 @@
 #' @example ../examples/qgrams.R
 #' @export
 qgrams <- function(..., .list=NULL,q=1L,useBytes=FALSE, useNames=!useBytes){
+ stopifnot(is.numeric(q), length(q)==1, !is.na(q), q>=0)
  q <- as.integer(q)
 
  if (!is.null(.list) && length(.list) == 0) .list=NULL
  L <- lapply(c(list(...),.list), as.character)
  if (length(L) == 0) return(array(dim=c(0,0)))
  L <- setnames(L)
+
+ if (q==0){
+ return( matrix( sapply(L,function(x) sum(x==""))
+ , ncol=1
+ , dimnames = list(names(L), NULL)) )
+ }
+
+
  L <- lapply(L,char2int)
 
  v <- .Call("R_get_qgrams",L,as.integer(q),PACKAGE="stringdist")

pkg/R/stringdist.R

@@ -63,7 +63,7 @@
 #' }
 #' Or use \code{citation('stringdist')} to get a bibtex item.
 #'
-#' @name stringdist-package
+#' @aliases stringdist-package
 #' @docType package
 #' @useDynLib stringdist, .registration=TRUE
 #' @importFrom parallel detectCores 

pkg/inst/tinytest/test_qgrams.R

@@ -9,7 +9,7 @@ options(sd_num_thread=2)
  expect_equivalent(qgrams(NA,q=1), matrix(0,nrow=1,ncol=0)) # skip all
  expect_equivalent(qgrams(c("a","ab"), q=2), as.matrix(table("ab"))) # skip q>nchar
  expect_equivalent(qgrams(c("a"),q=2), matrix(0,nrow=1,ncol=0)) # skip all
- expect_equivalent(qgrams(c(''),q=0), as.matrix(table(''))) # empty string, q=0
+ expect_equivalent(qgrams(c(''),q=0), matrix(table(''))) # empty string, q=0
 
 
 ## qgrams

pkg/src/qgram.c

@@ -546,7 +546,8 @@ SEXP R_get_qgrams(SEXP a, SEXP qq){
  for ( int i=0; i < nstr; ++i ){
  str = (unsigned int *) INTEGER(VECTOR_ELT(strlist,i));
  nchar = length(VECTOR_ELT(strlist,i));
- if ( str[0] == NA_INTEGER 
+ if ( nchar == 0
+ || str[0] == NA_INTEGER 
  || q > nchar
  || ( q == 0 && nchar > 0 )
  ){